Page 136 - Artificial Intelligence in the Age of Neural Networks and Brain Computing

4. Brain-Like AI Systems Based on SNN. NeuCube. 125

4. BRAIN-LIKE AI SYSTEMS BASED ON SNN. NEUCUBE.
DEEP LEARNING ALGORITHMS
4.1 BRAIN-LIKE AI SYSTEMS. NEUCUBE

Inspired by the structure and the functions of the human brain, a SNN learning
machine was developed, named NeuCube [106]. It was initially designed for spatio-
temporal brain data modeling, but then it was also used for climate data modeling
and stroke occurrence prediction and other applications [107].
The NeuCube framework is depicted in Fig. 6.10. It consists of the following
modules:
• Input information encoding module;
• 3D SNN reservoir module (SNNr);
• Output (e.g., classification) module;
• Gene regulatory network (GRN) module (optional);
• Optimization module (optional).

The input module transforms input data into trains of spikes. Spatiotemporal data
(such as EEG, climate, cybersecurity, financial, etc.) is entered after the encoding
into the main module dthe 3D SNN neurogenetic brain cube (NBC), or simply
the SNNcube. Input data is entered into pre-designated spatially distributed areas
of the SNNcube that correspond to the spatial location in the origin where data
was collected.

FIGURE 6.10
A block diagram of the brain-like SNN system NeuCube with evolving output module
[106].